Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A data-provider system may perform a dot-product operation using encrypted data, and a secure-processing component may decrypt and process that data using an activation function to predict an event. Multiple secure-processing components may be used to perform a multiplication operation using homomorphic encrypted data.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer-implemented method comprising: determining, by a data provider system, encrypted first input data and encrypted second input data, wherein at least one of the encrypted first input data and the encrypted second input data correspond to an event; determining, by the data provider system, an encrypted first random number and an encrypted second random number, wherein determining the encrypted first random number and the encrypted second random number comprises: determining a first random number and a second random number; encrypting, using private key data, the first random number, wherein the private key data corresponds to a private encryption key; and encrypting, using the private key data, the second random number; determining, by the data provider system, first data representing a result of a first homomorphic operation of the encrypted first input data, the encrypted second input data, the encrypted first random number, and the encrypted second random number, wherein determining the first data comprises: adding, using a homomorphic addition operation, the encrypted first input data and the encrypted first random number; and adding, using the homomorphic addition operation, the encrypted second input data and the encrypted second random number; sending, from the data provider system to a data processing system, the first data; decrypting, by the data processing system, the first data to determine a first number based at least in part on the encrypted first random number and a second number based at least in part on the encrypted second random number, wherein decrypting the first data corresponds to the private key data; multiplying, by the data processing system, the first number and the second number to determine second data; encrypting, by the data processing system, the second data to determine encrypted second data, wherein encrypting the second data corresponds to public key data, the public key data corresponding to a public encryption key; determining, by the data processing system using a second homomorphic operation, a product of the encrypted first input data and the encrypted second input data based at least in part on the encrypted second data; and determining a prediction of the event based at least in part on the product.
This invention relates to secure data processing for event prediction using homomorphic encryption. The system addresses the challenge of performing computations on encrypted data without exposing sensitive information, enabling privacy-preserving analysis of events. A data provider system encrypts input data and random numbers using a private key, then performs homomorphic additions to combine the encrypted inputs with the encrypted random numbers. The resulting encrypted data is sent to a data processing system, which decrypts it to recover the original values. The data processing system then multiplies these values, encrypts the product using a public key, and performs a homomorphic operation to compute the product of the original encrypted inputs. The final product is used to predict an event. The method ensures that intermediate computations remain encrypted, preserving data privacy while enabling secure event prediction. The approach leverages homomorphic encryption to perform mathematical operations on encrypted data without decryption, ensuring confidentiality throughout the process. The system is particularly useful in applications requiring secure analysis of sensitive data, such as financial forecasting, medical diagnostics, or risk assessment.
2. The computer-implemented method of claim 1 , further comprising: determining third data by adding the encrypted first input data and the encrypted first random number and by negating the encrypted second random number; determining fourth data by negation of the encrypted first random number and by adding the encrypted second input data and the encrypted second random number; and determining fifth data corresponding to the encrypted first random number and the encrypted second random number.
This invention relates to cryptographic data processing, specifically methods for securely combining encrypted data while preserving privacy. The problem addressed is the need to perform computations on encrypted data without revealing the underlying plaintext values, which is critical in privacy-preserving applications such as secure multi-party computation or confidential data analysis. The method involves processing two sets of encrypted input data and random numbers. First, encrypted first input data and an encrypted first random number are combined by addition, while an encrypted second random number is negated. This produces third data. Second, the encrypted first random number is negated, and encrypted second input data and the encrypted second random number are combined by addition, producing fourth data. Finally, fifth data is generated corresponding to the encrypted first and second random numbers. These operations allow secure computation on encrypted values without exposing the original data, enabling privacy-preserving arithmetic operations. The technique ensures that intermediate and final results remain encrypted, maintaining confidentiality throughout the process. This approach is useful in scenarios requiring secure aggregation of encrypted data, such as in privacy-preserving machine learning or confidential financial transactions.
3. The computer-implemented method of claim 1 , further comprising: receiving, by a first secure-processing component of the data processing system, the first data; and sending, by the first secure-processing component to a second secure-processing component of the data processing system, the encrypted second data.
A computer-implemented method enhances secure data processing in a system with multiple secure-processing components. The method addresses the challenge of securely transferring and processing sensitive data within a data processing system, ensuring confidentiality and integrity. The system includes at least two secure-processing components, each designed to handle data securely. The method involves receiving first data by a first secure-processing component, which then processes this data to generate encrypted second data. The encrypted second data is subsequently sent from the first secure-processing component to a second secure-processing component within the same system. This ensures that sensitive data remains protected during transfer and processing. The method may also include additional steps such as decrypting the encrypted second data by the second secure-processing component, further processing the decrypted data, and generating output data. The secure-processing components may be specialized hardware or software modules configured to enforce strict security policies, such as encryption, access control, and data isolation. The overall system ensures that data remains secure throughout its lifecycle, from initial receipt to final processing, mitigating risks of unauthorized access or tampering.
4. The computer-implemented method of claim 3 , further comprising: receiving, by the second secure-processing component, the encrypted second data; and sending, by the second secure-processing component to a third system, an indication of the prediction.
This invention relates to secure data processing in distributed systems, particularly for privacy-preserving machine learning or predictive analytics. The problem addressed is securely sharing and processing sensitive data across multiple systems while maintaining confidentiality and integrity. The method involves a first secure-processing component encrypting first data and sending it to a second secure-processing component. The second component decrypts the first data and generates a prediction based on it. The second component then encrypts second data related to this prediction and sends it to a third system. The third system receives this encrypted second data and sends an indication of the prediction to another system. The process ensures that sensitive data remains encrypted during transmission and processing, with decryption only occurring within secure processing environments. This approach enables secure collaboration between systems while preventing unauthorized access to raw or intermediate data. The invention is particularly useful in scenarios requiring strict data privacy, such as healthcare analytics, financial risk assessment, or federated learning systems where multiple parties contribute data without exposing it to each other. The method leverages cryptographic techniques to maintain data confidentiality throughout the processing pipeline.
5. The computer-implemented method of claim 1 , wherein the encrypted first input data and encrypted second input data correspond to operands of a dot-product operation.
The invention relates to secure computation techniques for performing mathematical operations on encrypted data. Specifically, it addresses the challenge of computing a dot-product operation between two sets of encrypted input data without decrypting the data, thereby preserving data privacy. The method involves processing encrypted first input data and encrypted second input data, where these encrypted inputs represent operands of a dot-product operation. The dot-product operation is a fundamental computation in machine learning, cryptography, and data analysis, where the sum of element-wise products of two vectors is calculated. By performing this operation directly on encrypted data, the method ensures that sensitive information remains confidential throughout the computation. The technique leverages cryptographic protocols to enable secure multiplication and addition of encrypted values, allowing the dot-product to be computed without exposing the underlying plaintext data. This approach is particularly useful in scenarios where data privacy is critical, such as in federated learning, secure multi-party computation, or confidential cloud computing. The method ensures that intermediate and final results remain encrypted, preventing unauthorized access to the original data while still producing accurate computational results.
6. The computer-implemented method of claim 1 , further comprising: determining that the encrypted first input data corresponds to a first scale; determining that the encrypted second input data corresponds to a second scale different from the first scale; and modifying the encrypted first input data to correspond to the second scale.
This invention relates to secure data processing, specifically handling encrypted input data with different scales. The problem addressed is ensuring accurate and secure computations when encrypted data from different sources or systems have incompatible scales, which can lead to errors or incorrect results in encrypted computations. The method involves receiving encrypted first and second input data, where the data is encrypted to preserve privacy during processing. The system determines that the encrypted first input data corresponds to a first scale (e.g., a measurement unit or range) and that the encrypted second input data corresponds to a second, different scale. To enable accurate computations, the method modifies the encrypted first input data to match the second scale while maintaining the encryption. This ensures that the data remains secure and usable in subsequent encrypted operations, such as comparisons, aggregations, or other computations. The modification process may involve scaling transformations, unit conversions, or other adjustments applied directly to the encrypted data without decrypting it. This approach allows secure and seamless integration of data from disparate sources, improving the reliability of encrypted data processing systems. The invention is particularly useful in privacy-preserving applications like secure analytics, financial transactions, or healthcare data processing.
7. The computer-implemented method of claim 1 , wherein the encrypted first input data and encrypted second input data correspond to elements of a vector.
This invention relates to a computer-implemented method for processing encrypted data in a vectorized format. The method addresses the challenge of securely handling sensitive information while performing computations, particularly in applications like secure multi-party computation or privacy-preserving data analysis. The technique involves encrypting input data elements and organizing them into a vector structure, enabling efficient and secure operations on the encrypted data without exposing the underlying plaintext values. The method builds on a foundational process that encrypts first and second input data elements, ensuring confidentiality during storage or transmission. These encrypted elements are then mapped to corresponding positions within a vector, allowing for batch processing or parallel computations. The vectorized approach enhances performance by enabling simultaneous operations on multiple encrypted data points, which is particularly useful in cryptographic protocols requiring large-scale data manipulation. The encrypted vector elements may be processed using homomorphic encryption techniques, allowing computations to be performed directly on the encrypted data while preserving the integrity of the results. This approach ensures that sensitive information remains protected throughout the computation, addressing concerns related to data privacy and security in distributed or untrusted environments. The method is applicable in fields such as secure cloud computing, financial transactions, and healthcare data analysis, where maintaining data confidentiality is critical.
8. A system comprising: at least one processor; and at least one memory including instructions that, when executed by the at least one processor, cause the system to: determine, by a data provider system, encrypted first input data and encrypted second input data, wherein at least one of the encrypted first input data and the encrypted second input data correspond to an event; determine, by the data provider system, an encrypted first random number and an encrypted second random number, wherein determining the encrypted first random number and the encrypted second random number, the at least one memory further includes instructions, that, when executed by the at least one processor, further cause the system to: determine a first random number and a second random number; encrypt, using private key data, the first random number, wherein the private key data corresponds to a private encryption key; and encrypt, using the private key data, the second random number; determine, by the data provider system, first data representing a result of a first homomorphic operation of the encrypted first input data, the encrypted second input data, the encrypted first random number, and the encrypted second random number, wherein the at least one memory further includes instructions, that, when executed by the at least one processor, further cause the system to: add, using a homomorphic addition operation, the encrypted first input data and the encrypted first random number; and add, using the homomorphic addition operation, the encrypted second input data and the encrypted second random number; send, from the data provider system to a data processing system, the first data; decrypt, by the data processing system, the first data to determine a first number based at least in part on the encrypted first random number and a second number based at least in part on the encrypted second random number, wherein decrypting the first data corresponds to the private key data; multiply, by the data processing system, the first number and the second number to determine second data; encrypt, by the data processing system, the second data to determine encrypted second data, wherein encrypting the second data corresponds to public key data, the public key data corresponding to a public encryption key; determine, by the data processing system using a second homomorphic operation, a product of the encrypted first input data and the encrypted second input data based at least in part on the encrypted second data; and determine a prediction of the event based at least in part on the product.
This system enables secure computation of event predictions using homomorphic encryption to protect sensitive input data. The system involves a data provider system and a data processing system that collaborate to compute a product of two encrypted input values without exposing the original data. The data provider system generates encrypted random numbers and performs homomorphic additions with the encrypted input data, producing intermediate results. These results are sent to the data processing system, which decrypts them to obtain two numbers derived from the random values. The data processing system then multiplies these numbers to produce a new encrypted value, which is sent back to the data provider system. The data provider system uses this value in a second homomorphic operation to compute the product of the original encrypted inputs. The final product is used to generate a prediction about an event represented by the input data. The system ensures that neither party can access the other's raw data, preserving privacy while enabling secure computation. This approach is useful for applications requiring confidential data analysis, such as financial forecasting, medical diagnostics, or risk assessment, where sensitive information must remain protected throughout the computation process.
9. The system of claim 8 , wherein the at least one memory further includes instructions, that, when executed by the at least one processor, further cause the system to: determine third data by adding the encrypted first input data and the encrypted first random number and by negating the encrypted second random number; determine fourth data by negation of the encrypted first random number and by adding the encrypted second input data and the encrypted second random number; and determine fifth data corresponding to the encrypted first random number and the encrypted second random number.
This invention relates to a cryptographic system for secure data processing, particularly in scenarios requiring privacy-preserving computations. The system addresses the challenge of performing operations on encrypted data without revealing the underlying plaintext values, which is critical for applications like secure multi-party computation, confidential cloud computing, or privacy-preserving machine learning. The system includes at least one processor and memory storing instructions for processing encrypted data. The memory contains encrypted first and second input data, as well as encrypted first and second random numbers. The processor executes instructions to generate intermediate encrypted results by performing arithmetic operations on these values. Specifically, the system computes third data by adding the encrypted first input data and the encrypted first random number, then subtracting the encrypted second random number. It also computes fourth data by negating the encrypted first random number and adding the encrypted second input data and the encrypted second random number. Additionally, the system determines fifth data representing the encrypted first and second random numbers. These operations enable secure transformations of encrypted data while preserving confidentiality, allowing further cryptographic or computational steps to be performed without exposing sensitive information. The approach leverages homomorphic properties of encryption to maintain data privacy throughout processing.
10. The system of claim 8 , wherein the at least one memory further includes instructions, that, when executed by the at least one processor, further cause the system to: receive, by a first secure-processing component of the data processing system, the first data; and send, by the first secure-processing component to a second secure-processing component of the data processing system, the encrypted second data.
This invention relates to a data processing system with secure components for handling sensitive data. The system addresses the challenge of securely transferring and processing data within a computing environment, particularly where data must be encrypted and transmitted between secure processing units to prevent unauthorized access. The system includes at least one processor and memory storing instructions that, when executed, enable secure data handling. A first secure-processing component receives initial data, which is then encrypted into a second form. This encrypted data is transmitted to a second secure-processing component within the same system. The secure-processing components are specialized hardware or software modules designed to isolate and protect sensitive operations, ensuring data integrity and confidentiality during transfer. The system may also include additional components for further processing or storage of the encrypted data, depending on the application. This approach enhances security by restricting access to encrypted data only to authorized secure-processing components, reducing exposure to potential breaches. The system is particularly useful in environments requiring high levels of data protection, such as financial transactions, healthcare records, or government communications. The invention ensures that sensitive information remains encrypted during transit between secure modules, mitigating risks associated with interception or tampering.
11. The system of claim 10 , wherein the at least one memory further includes instructions, that, when executed by the at least one processor, further cause the system to: receive, by the second secure-processing component, the encrypted second data; and send, by the second secure-processing component to a third system, an indication of the prediction.
The system relates to secure data processing and prediction systems, particularly for handling sensitive or confidential data while ensuring privacy and security. The problem addressed is the need to process and analyze data in a secure manner, preventing unauthorized access or exposure of sensitive information during processing, storage, or transmission. The system includes multiple secure-processing components that operate in isolated environments to protect data integrity and confidentiality. One component encrypts data before processing, while another decrypts and processes the data securely. The system also generates predictions based on the processed data and communicates these predictions to external systems without exposing the underlying raw data. The secure-processing components are designed to prevent unauthorized access to the data, ensuring that sensitive information remains protected throughout the entire processing pipeline. The system is particularly useful in applications where data privacy and security are critical, such as healthcare, finance, or government sectors. The invention ensures that data is processed in a way that complies with regulatory requirements and industry standards for data protection.
12. The system of claim 8 , wherein the encrypted first input data and encrypted second input data correspond to operands of a dot-product operation.
The invention relates to a cryptographic system designed to perform secure computations on encrypted data. The system addresses the challenge of enabling computations on encrypted data without exposing the underlying plaintext, which is critical for privacy-preserving applications such as secure machine learning, financial transactions, and confidential data processing. The system includes a processing unit configured to receive encrypted first input data and encrypted second input data. These encrypted inputs correspond to operands of a dot-product operation, a fundamental computation in linear algebra and machine learning. The processing unit performs a dot-product operation on the encrypted inputs, producing an encrypted result that maintains the confidentiality of the original data. This allows secure evaluation of mathematical operations without decrypting the inputs, preserving privacy throughout the computation. The system may also include a key management module to handle cryptographic keys used for encryption and decryption, ensuring secure access and processing. The processing unit may further include specialized hardware or software components optimized for performing dot-product operations on encrypted data efficiently, such as homomorphic encryption techniques that allow computations on ciphertexts to yield encrypted results. By enabling secure dot-product operations on encrypted data, the system supports privacy-preserving applications where sensitive data must remain confidential during processing. This is particularly useful in scenarios like secure cloud computing, federated learning, and confidential data analysis, where data owners cannot or do not want to share their raw data.
13. The system of claim 8 , wherein the at least one memory further includes instructions, that, when executed by the at least one processor, further cause the system to: determine that the encrypted first input data corresponds to a first scale; determine that the encrypted second input data corresponds to a second scale different from the first scale; and modify the encrypted first input data to correspond to the second scale.
This invention relates to a data processing system that handles encrypted input data with different scales. The system includes at least one processor and at least one memory storing instructions. The system processes encrypted first and second input data, where the first input data is encrypted and corresponds to a first scale, and the second input data is encrypted and corresponds to a second scale different from the first. The system determines the scales of the encrypted input data and modifies the encrypted first input data to match the second scale, ensuring consistency in scale between the two encrypted datasets. This allows for accurate processing of encrypted data without decrypting it, maintaining data security while enabling operations that require uniform scaling. The system may be part of a larger data processing framework, such as a secure computing environment or a privacy-preserving analytics platform, where encrypted data must be manipulated without exposure. The invention addresses the challenge of performing operations on encrypted data with varying scales, ensuring compatibility and correctness in computations while preserving confidentiality.
14. The system of claim 8 , wherein the encrypted first input data and encrypted second input data correspond to elements of a vector.
A system for secure data processing involves encrypting input data elements to form a vector of encrypted values. The system processes these encrypted elements while preserving their mathematical relationships, enabling computations on encrypted data without exposing the underlying plaintext values. This approach addresses the challenge of performing secure computations on sensitive data, such as in privacy-preserving machine learning or financial transactions, where data confidentiality must be maintained. The encrypted input data elements are structured as a vector, allowing operations like linear transformations or dot products to be performed directly on the encrypted values. The system ensures that intermediate and final results remain encrypted, preventing unauthorized access to the original data. By maintaining the vector structure, the system supports efficient and scalable computations while adhering to cryptographic security principles. This method is particularly useful in applications requiring secure data analysis, such as federated learning or confidential cloud computing, where multiple parties collaborate without sharing raw data. The system leverages cryptographic techniques to protect data integrity and confidentiality throughout the processing pipeline.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
May 25, 2021
March 22, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.